The Neglected Response Variable

2005 ◽  
pp. 61-82
Keyword(s):  
2020 ◽  
Vol 8 (1) ◽  
pp. 157-171 ◽  
Author(s):  
Himchan Jeong ◽  
Emiliano A. Valdez

AbstractFor observations over a period of time, Bayesian credibility premium may be used to predict the value of a response variable for a subject, given previously observed values. In this article, we formulate Bayesian credibility premium under a change of probability measure within the copula framework. Such reformulation is demonstrated using the multivariate generalized beta of the second kind (GB2) distribution. Within this family of GB2 copulas, we are able to derive explicit form of Bayesian credibility premium. Numerical illustrations show the application of these estimators in determining experience-rated insurance premium. We consider generalized Pareto as a special case.


2020 ◽  
Vol 8 (1) ◽  
pp. 54-69
Author(s):  
Peter B. Gilbert ◽  
Bryan S. Blette ◽  
Bryan E. Shepherd ◽  
Michael G. Hudgens

AbstractWhile the HVTN 505 trial showed no overall efficacy of the tested vaccine to prevent HIV infection over placebo, markers measuring immune response to vaccination were strongly correlated with infection. This finding generated the hypothesis that some marker-defined vaccinated subgroups were partially protected whereas others had their risk increased. This hypothesis can be assessed using the principal stratification framework (Frangakis and Rubin, 2002) for studying treatment effect modification by an intermediate response variable, using methods in the sub-field of principal surrogate (PS) analysis that studies multiple principal strata. Unfortunately, available methods for PS analysis require an augmented study design not available in HVTN 505, and make untestable structural risk assumptions, motivating a need for more robust PS methods. Fortunately, another sub-field of principal stratification, survivor average causal effect (SACE) analysis (Rubin, 2006) – which studies effects in a single principal stratum – provides many methods not requiring an augmented design and making fewer assumptions. We show how, for a binary intermediate response variable, methods developed for SACE analysis can be adapted to PS analysis, providing new and more robust PS methods. Application to HVTN 505 supports that the vaccine partially protected individuals with vaccine-induced T-cells expressing certain combinations of functions.


Micromachines ◽  
2019 ◽  
Vol 10 (5) ◽  
pp. 335 ◽  
Author(s):  
Antonio Luca ◽  
Oltmann Riemer

Microinjection moulding has been developed to fulfil the needs of mass production of micro components in different fields. A challenge of this technology lies in the downscaling of micro components, which leads to faster solidification of the polymeric material and a narrower process window. Moreover, the small cavity dimensions represent a limit for process monitoring due to the inability to install in-cavity sensors. Therefore, new solutions must be found. In this study, the downscaling effect was investigated by means of three spiral geometries with different cross sections, considering the achievable flow length as a response variable. Process indicators, called “process fingerprints”, were defined to monitor the process in-line. In the first stage, a relationship between the achievable flow length and the process parameters, as well as between the process fingerprints and the process parameters, was established. Subsequently, a correlation analysis was carried out to find the process indicators that are mostly related to the achievable flow length.


2012 ◽  
Vol 57 (1) ◽  
Author(s):  
SEYED EHSAN SAFFAR ◽  
ROBIAH ADNAN ◽  
WILLIAM GREENE

A Poisson model typically is assumed for count data. In many cases, there are many zeros in the dependent variable and because of these many zeros, the mean and the variance values of the dependent variable are not the same as before. In fact, the variance value of the dependent variable will be much more than the mean value of the dependent variable and this is called over–dispersion. Therefore, Poisson model is not suitable anymore for this kind of data because of too many zeros. Thus, it is suggested to use a hurdle Poisson regression model to overcome over–dispersion problem. Furthermore, the response variable in such cases is censored for some values. In this paper, a censored hurdle Poisson regression model is introduced on count data with many zeros. In this model, we consider a response variable and one or more than one explanatory variables. The estimation of regression parameters using the maximum likelihood method is discussed and the goodness–of–fit for the regression model is examined. We study the effects of right censoring on estimated parameters and their standard errors via an example.


2018 ◽  
Author(s):  
◽  
Li Chen

[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT AUTHOR'S REQUEST.] Longitudinal data contain repeated measurements of variables on the same experimental subject. It is often of interest to analyze the relationship between these variables. Typically, there is one or several longitudinal covariates and a response variable that can be either longitudinal or time to an event. Regression models can be employed to analyze these relationships. Ideally, longitudinal variables should be continuously monitored and their complete trajectories along the time are observed. Practically, however, this is unrealistic, either economically or methodologically. Often one only obtains so called sparse longitudinal data, where variables are intermittently observed at relatively sparse time points within the period of study. Such sparse longitudinal data give rise to an issue for the analysis of the response of time to an event, where survival analysis is typically implemented, e.g. the Cox model or additive hazards model. In both models, the values of covariates of all subjects at risk are needed in order to calculate the partial likelihood. But in the case of sparse longitudinal data, the availability of these observations may not be satis fied. Moreover, if the response variable is also longitudinal, it is possible that the response and covariates are not observed altogether, or at least not close to each other enough to be considered as observed simultaneously. Although a wealth of studies have been dedicated to longitudinal data analysis, very few of them have seriously considered and rigorously studied the situation aforementioned. In this dissertation, we discuss the regression analysis of longitudinal cavities with censored and longitudinal outcome. To be specific, Chapter 2 targets the additive hazards models with sparse longitudinal covariates, Chapter 3 studies the partially linear models with longitudinal covariates and response observed at mismatched time points, also known as asynchronous longitudinal data, and Chapter 4 explores longitudinal data with more complex structures with linear models. Kernel weighting technique is the key idea to all the stated researches. Estimators are derived based on kernel weighting technique and their asymptotical properties were rigorously examined, along with simulation studies for their fi nite sample performance, and illustrations using real data sets.


2016 ◽  
Vol 5 (4) ◽  
pp. 133
Author(s):  
NI PUTU PREMA DEWANTI ◽  
MADE SUSILAWATI ◽  
I GUSTI AYU MADE SRINADI

Poisson regression is a nonlinear regression which is often used for count data and has equidispersion assumption (variance value equal to mean value). However in practice, equidispersion assumption is often violated. One of it violations is overdispersion (variance value greater than the mean value). One of the causes of overdipersion is excessive number of zero values on the response variable (excess zeros). There are many methods to handle overdispersion because of excess zeros. Two of them are Zero Inflated Poisson (ZIP) regression and Zero Inflated Negative Binomial (ZINB) regression. The purpose of this research is to determine which regression models is better in handling overdispersion data. The data that can be analyzed using the ZIP and ZINB regression is maternal mortality rate in the Province of Bali. Maternal mortality rate data has proportion of zeros value more than 50% on the response variable.  In this research, ZINB regression better than ZIP regression for modeling maternal mortality rate. The independent variable that affects the number of maternal mortality rate in the Province of Bali  is the percentage of mothers who carry a pregnancy visit, with ZINB regression models and . 


2018 ◽  
Vol 22 (8) ◽  
pp. 4593-4604 ◽  
Author(s):  
Yongqiang Zhang ◽  
David Post

Abstract. Gap-filling streamflow data is a critical step for most hydrological studies, such as streamflow trend, flood, and drought analysis and hydrological response variable estimates and predictions. However, there is a lack of quantitative evaluation of the gap-filled data accuracy in most hydrological studies. Here we show that when the missing data rate is less than 10 %, the gap-filled streamflow data obtained using calibrated hydrological models perform almost the same as the benchmark data (less than 1 % missing) when estimating annual trends for 217 unregulated catchments widely spread across Australia. Furthermore, the relative streamflow trend bias caused by the gap filling is not very large in very dry catchments where the hydrological model calibration is normally poor. Our results clearly demonstrate that the gap filling using hydrological modelling has little impact on the estimation of annual streamflow and its trends.


Sign in / Sign up

Export Citation Format

Share Document